Project 1: Time Domian Electron Paramagnetic Resonance Imaging: Instrumentation: Programmable Timing Unit: The time resolution of the radiofrequency pulses needed in EPR Imaging at any given frequency requires time resolution in nanosecond range unlike MRI experiments at the same frequency. This is because of the faster spin dynamics of paramagnetic spin systems compared to nuclear spin dynamics. The timing unit to manage the RF circuitry is not available commercially. We have designed, tested, and integrated a novel programmable timing unit with nanosecond time resolution which controls all the RF modules in the spectrometer using a new approach in RF electronics utilizing LabView technology. Integrating this unit resulted in simplifying the spectrometer operation significantly so that general purpose users can use the scanner without RF engineering expertise. In addition, this addition to the spectrometer made the unit less bulky and decreased the cost of the second spectrometer being built for the newly purchased open magnet system. Open Magnet System: The currently used magnet for all the studies was purchased in 1994 and has provided reliable performance. However, the small bore size and design makes in vivo experiments difficult in terms of housing the animal and have the anesthesia, iv lines etc and the air handling in the bore to maintain the core body temperature at 37 C. To overcome these difficulties, we have procured and installed an open magnet system with access in both directions in the horizontal plane (X- and Y-directions) so that in vivo experiments can be performed with ease. The magnet and gradient coils have been calibrated with the corresponding amplifiers and the control software completed. The RF chain is integrated and the system is ready for operation. Gantry for EPRI-MRI combined imaging: Since EPR based pO2 images lack anatomic information, interpreting the oxygen maps is often equivocal. To overcome this limitation, we have developed strategies for co-registering the EPR Images with anatomic images from MRI. The frequency of operation in EPRI of 300 MHz is similar to that of an MRI scanner operating at 7 T. Therefore a common resonator and gantry were designed, tested, optimized for operation in both modalities and used for sequential imaging of the object with EPRI and RI without moving the object. This made it possible to interpret the oxygen maps more reliable utilizing the anatomic guidance. In Vivo EPRI and MRI Co-Imaging: EPR imaging can visualize the distribution of oxygen concentration in tumor, though it doesnt provide anatomical information. The combined system of EPRI/MRI makes it possible to know the exact location of hypoxic core in the tumor anatomy, and also multi-functional analysis of tumor physiology combining the oxygen status and other functions by MRI including blood volume, blood flow, water diffusion etc. 300 MHz pulsed EPR/oxygen imaging system was constructed in which the RF coil and gantry were designed to also be used for MRI. The mouse to be measured is transferred between EPR and MRI magnets without removing the mouse from the coil, making reliable and simple co-registration of oxygen map by EPRI and anatomy by MRI possible. Using this EPR/MRI co-registration system, the relationship among tumor oxygen status, other parameters including blood perfusion, water diffusion which inversely related with tumor cellularity, MR spectroscopy, and resulting output of radiation therapy is underway. Image Reconstruction: The currently used EPRI image reconstruction algorithms are adapted on a Pentium PC. The current situation is that the data sets are to calculate the pO2 images with the required spatial and physiological resolutions. It takes 2 hours computational time to reorganize the acquired data and an additional 30 minutes to calculate the pO2 values in each voxel. This imposes enormous burden to process routinely collected image data from in vivo experiments. To reduce the image reconstruction time, we have acquired a Linux-based four dual-core cluster CPU with a 2 TB memory. Currently, whenever the phantom/animal in-vivo oximetry experiment is completed, the 3-D oximetry data sets that are collected in the collection machine are transferred to the Parallel Server via network I/O. A unique index file is created during every experiment and used by the server to construct PBS (Parallel Batch Server) job. The FIDs are reduced by automatically running a multi-threaded C program by the four dual-core processors of the machine. A graphical user interface (GUI) developed using custom parallel Matlab, uses the reduced data sets to view the mesh plots and to reconstruct the concentration images and oxygen images depending upon the input parameters provided by the user. The time complexity has now drastically reduced in the order of few minutes. A Windows PC client to Linux server connection via Samba (Session Message Block) connectivity and SSH (Secure Shell) enables any authorized users to run the GUI